Taking Action to Reduce Tobacco Use (1998)

differentiation in function. As a universal program, it is logical to shift responsibility to the nation's public health agency, CDC. This would entail a significant scaling-up of current efforts. Yet expansion of the CDC effort should not detract from further progress. Tobacco control efforts will be expanding in Alaska, Arizona, Hawaii, Maine, Oregon, Utah, Wisconsin, and other states. The ASSIST program has shown that a more intense intervention produces results, but it does not clearly show which elements are most powerful, and ASSIST states could not test counteradvertising, public education, or other interventions at the level possible in California and Massachusetts. An expanded commitment to tobacco control increases the importance of knowing which interventions matter most, requiring demonstrations at sufficient dose and duration to enable credible evaluation. That is, many questions remain for research, high-intensity demonstrations, and rigorous program evaluation. This is no time to retrench from research or to stop testing even larger-scale demonstrations. The goal for the next decade should be to achieve tobacco control rivaling that in California and Massachusetts, through the use of a combination of state, local, and federal funds and any payments resulting from a national settlement, and to improve the programs in all states through research, demonstrations, and program evaluation.

The most successful efforts to curb tobacco use have grown from mobilization at the state and community levels and have entailed collaborations among private health organizations and federal, state, and local governments. Federal legislation currently prohibits state and local governments from regulating any form of advertising and promotional activities based on smoking and health, even if the activity occurs exclusively within states' jurisdictional borders. The opportunity for innovation at the state and local levels would be enhanced if this federal impediment were removed, as recommended in the 1994 IOM report.

No matter what the final settlements about federal policy prove to be, federal action alone is insufficient to achieve tobacco control. Regardless of which tobacco control measures are put in place, both local and national capacities must be in place to translate goals into actions, to measure progress, and to provide feedback for subsequent policy decisions. The massive national effort that went into Healthy People 2000 is a good place to start a discussion of public health goals.²⁴Healthy People 2000 devoted a chapter to tobacco use, specifying national goals on a timetable; the goals for reducing tobacco use among youths, however, are further from achievement than they were even a few years ago.

A recent IOM report, Improving Health in the Community,²⁵ proposes ways of bridging the gap between national goals and community action and identifying specific organizations or groups at the community level accountable for making progress toward those goals. The general approach discussed in the report is embodied in the Community Health Improvement Process (CHIP). CHIP applies the tools used to monitor performance. To promote the achievement of health improvement goals, the approach entails (a) specifying goals, (b) developing a strategy to achieve them, (c) identifying and implementing local interventions that can be monitored by quantitative indicators, and then (d) collecting and assessing performance data for the specific accountable entities at the community level to evaluate the effectiveness of the intervention strategy and the contributions of the specific accountable entities in the community. The accountable entities will vary among communities, as will the relative priorities of the goals and the resources available to attain those goals.

The CHIP approach might be adapted for use at the national level as well. National health improvement goals might be translated into intervention strategies that federal agencies and a variety of national organizations might be expected to act on and for which performance indicators might be developed and monitored. Such activities at the national level might help shape related efforts at the state and local levels. The approach is described at greater length, with its theoretical underpinning and with specific consideration of tobacco control issues, in a separate background paper prepared for the board by Michael Stoto and Jane Durch of IOM (available on-line at http://www2.nas.edu/cancerbd/226e.html).

Three elements are central to the IOM CHIP model: (1) a broad view of health as a product of the interaction of many factors; (2) recognition that protecting and improving health is a shared responsibility of many entities, each of which needs to be accountable for its activities; and (3) a performance monitoring framework, in which sets of actionable measures are tied to specific entities that can help to ensure the necessary accountability. Accountability is a concern because responsibility diffused among many entities can easily be ignored or abandoned. This approach also points to the importance of interventions that focus not only on individuals but also on collective actions such as the adoption of policies to limit smoking in workplaces or restaurants or the enforcement of laws prohibiting the sale of tobacco products to persons under age 18. In extending the CHIP concept to the national level, there may be opportunities to focus on interventions that target communities and organizations rather than individuals.

A well-chosen set of performance measures could help to reinforce a balanced national and local tobacco control effort. The CHIP approach calls for the use of sets of indicators to make meaningful assessments of overall performance because health issues have many dimensions and can be addressed by various sectors. These sets of indicators should cover critical features of a health improvement effort. They should, however, remain selective; too many details can obscure the broader picture. Indicators must be carefully selected to provide insight into the progress that has been achieved. For an issue such as tobacco control, for which changes in health outcomes such as a reduction in lung cancer deaths will not be observable in the near term, the set of indicators should balance measures of shorter-term gains (e.g., reductions in smoking prevalence or sales of tobacco products to minors) and more fundamental longer-term changes in health (e.g., reductions

in the incidence of lung cancer or lung cancer mortality). The approach is perhaps best explained by way of an illustration, showing some possible national and local performance measures (see table, page 14).

The objectives in Healthy People 2000 are an essential reference point for developing performance measures at the national or the community level. The proposed FDA regulations that mandate restrictions on the sale of tobacco products to youths will be an appropriate focus for performance monitoring, even if other FDA provisions are not implemented. The proposed settlement also stipulates specific goals for reductions in the rate of smoking among youths that could guide the development of performance indicators. Likewise, the recommendations made by the Koop-Kessler Advisory Committee on Tobacco and Public Health suggest a variety of performance measures that might be used to monitor progress.

Without an overall monitoring plan, there is real danger of focusing on one or a few measures to the exclusion of others. The proposed settlement, for example, stipulates extra payments from tobacco firms if smoking measures for youths do not improve. The trigger is the level of daily smoking reported in surveys of youths conducted by the University of Michigan. This measure, part of the larger Monitoring the Future study, could change, even without real changes in the rate of initiation.²⁶ Initiation of tobacco use often begins well before high school and often progresses over two to three years. A more sensitive indicator, such as whether tobacco products were used in the previous 30 days, could also be used (with differential weight) earlier in the process to monitor tobacco use among those in younger age groups. These measures hinge on how students, excluding those not attending school, recall and report their smoking behavior. These measures are useful, but they should be balanced with other independent measures, such as smoking rates among those in the youngest age categories of the National Health Interview Survey, as suggested in the proposed performance measures (see tables). Otherwise, the high stakes create strong incentives to "game" a single indicator without changing the underlying behavior. Self-report measures are notoriously subject to interpretation and recall bias. The point is not that levels of smoking among youths reported in one survey or another are inaccurate or misleading—all measures are limited in one way or another—but rather that caution must be taken against relying on a single measure to monitor performance.

Some fraction of any additional revenues that become available should be devoted to getting better measures more frequently and to making them publicly available. Credible brand-specific penalties or tax increases as recommended above, for example, would require firm data the government could rely on when assessing the payments or taxes. At present the public health monitoring process is slow and coarse compared to the data available to individual tobacco firms.

Reducing tobacco use will require a broader set of measures, reflecting not only the initiation of tobacco use among youths but also the cessation of tobacco use among individuals in all age groups and the attainment of policy objectives. It will also require assessment at the community level as well as in national measures. The performance measures in the table mentioned above are merely first rough cuts to illustrate the possibilities of performance monitoring. The specific measures and objectives can surely be refined, but any long-term strategy for tobacco control must confront the problem: Broader and more robust ways of promoting and monitoring progress and evaluating the relative success of different interventions in diverse communities are needed.